QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Vol. 18, Issue 7, 2542-2560, July 2007

This Article Abstract Full Text (PDF) All Versions of this Article: E06-09-0822v1 18/7/2542    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Gao, X.-D. Articles by Bi, E. Search for Related Content PubMed PubMed Citation Articles by Gao, X.-D. Articles by Bi, E.

Sequential and Distinct Roles of the Cadherin Domain-containing Protein Axl2p in Cell Polarization in Yeast Cell Cycle

Xiang-Dong Gao^*,, Lauren M. Sperber^*, Steven A. Kane^*, Zongtian Tong^*, Amy Hin Yan Tong, Charles Boone, and Erfei Bi^*

*Department of Cell and Developmental Biology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104-6058; and Banting and Best Department of Medical Research, University of Toronto, Toronto, Ontario M5S 3E1, Canada

Submitted September 18, 2006; Revised April 10, 2007; Accepted April 18, 2007
Monitoring Editor: Tim Stearns

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Polarization of cell growth along a defined axis is essential for the generation of cell and tissue polarity. In the budding yeast Saccharomyces cerevisiae, Axl2p plays an essential role in polarity-axis determination, or more specifically, axial budding in MATa or cells. Axl2p is a type I membrane glycoprotein containing four cadherin-like motifs in its extracellular domain. However, it is not known when and how Axl2p functions together with other components of the axial landmark, such as Bud3p and Bud4p, to direct axial budding. Here, we show that the recruitment of Axl2p to the bud neck after S/G2 phase of the cell cycle depends on Bud3p and Bud4p. This recruitment is mediated via an interaction between Bud4p and the central region of the Axl2p cytoplasmic tail. This region of Axl2p, together with its N-terminal region and its transmembrane domain, is sufficient for axial budding. In addition, our work demonstrates a previously unappreciated role for Axl2p. Axl2p interacts with Cdc42p and other polarity-establishment proteins, and it regulates septin organization in late G1 independently of its role in polarity-axis determination. Together, these results suggest that Axl2p plays sequential and distinct roles in the regulation of cellular morphogenesis in yeast cell cycle.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Development of cell polarity is essential for the formation of diverse cell types in multicellular organisms, and it is critical for carrying out a number of cellular functions, including neuronal transmission, nutrient transport across epithelia, chemotaxis, and asymmetric cell division. Studies from various model systems indicate that the overall design and the core mechanisms underlying polarity establishment are conserved from yeast to humans, which involves a hierarchy of temporally and spatially coordinated events, including selection of a polarity-axis, cytoskeletal polarization, and cytoskeleton-guided protein trafficking (Drubin and Nelson, 1996; Nelson, 2003). In the budding yeast Saccharomyces cerevisiae, polarity-axis determination or bud-site selection differs in different cell types (Hicks et al., 1977; Chant and Pringle, 1995). Haploid MATa or cells bud in an axial pattern, i.e., new buds are formed adjacent to the site of previous cell division. Diploid MATa/ cells bud in a bipolar pattern, i.e., daughter cells bud at the distal pole, which is opposite to the site of cell division, whereas mother cells bud at either the distal or the proximal pole. The current thought on bud-site selection and polarized growth in S. cerevisiae is that an axial or a bipolar landmark in the cell cortex, which is assembled in previous cell cycle(s), is recognized by the Rsr1p (also known as Bud1p) GTPase module, and the positional information is relayed to polarity-establishment machinery, which is centered on the conserved Cdc42p GTPase module, resulting in the recruitment and the activation of Cdc42p at the chosen site (Chant, 1999; Casamayor and Snyder, 2002; Park and Bi, 2007). Activated Cdc42p polarizes the actin cytoskeleton, including actin patches and actin cables, which are involved in endocytosis and exocytosis, respectively; and polarized exocytosis results in bud formation (Moseley and Goode, 2006; Park and Bi, 2007).

Genetic analysis indicates that the axial landmark consists of Axl1p, Axl2p (also known as Bud10p), Bud3p, Bud4p, and the septins (Cdc3p, Cdc10p, Cdc11p, Cdc12p, and Shs1p/Sep7p) (Flescher et al., 1993; Fujita et al., 1994; Adames et al., 1995; Chant et al., 1995; Halme et al., 1996; Roemer et al., 1996; Sanders and Herskowitz, 1996). Null or conditional-lethal mutations in each of these genes result in bipolar budding in MATa or cells, but they have no effect on the budding pattern of MATa/ cells. The bipolar landmark consists of at least Bud8p, Bud9p, Rax1p, and Rax2p (Chen et al., 2000; Harkins et al., 2001; Fujita et al., 2004; Kang et al., 2004a). Null mutations in any of these genes result in unipolar (for BUD8 and BUD9) or nonrandom (RAX1 and RAX2) budding in MATa/ cells, but they have no effect on the axial budding of MATa or cells. The axial- or bipolar-landmark proteins are thought to recruit Bud5p, the guanine nucleotide-exchange factor (GEF) for Rsr1p, and also Bud2p, the GTPase-activating protein (GAP) for Rsr1p, to the cortical site, which activates the GTPase cycling of Rsr1p (Park et al., 1999; Kang et al., 2001, 2004b; Marston et al., 2001). Null mutations in RSR1, BUD2, or BUD5 result in random budding in all cell types. Rsr1p-GTP binds to Cdc24p, the GEF for Cdc42p, whereas Rsr1p-GDP binds to Bem1p, a polarity-establishment protein that also binds to Cdc24p and Cdc42p-GTP (Zheng et al., 1995; Park et al., 1997; Gulli et al., 2000; Bose et al., 2001; Kozminski et al., 2003). These molecular interactions constitute positive-feedback loops to localize, activate and amplify Cdc42p at the chosen site. As a consequence, a new bud site is assembled and is marked by polarized actin cytoskeleton and a septin ring.

Axl2p plays a key role in axial budding. First, the budding pattern of axl1, bud3, or bud4 cells is reverted to axial budding by rax1 (Lord et al., 2002; Fujita et al., 2004); in contrast, axl2 rax1 cells bud randomly (Fujita et al., 2004), suggesting that Axl2p is the true axial landmark whereas the functions of Axl1p, Bud3p, and Bud4p in axial budding may be more regulatory in nature. Second, Axl2p is thought to recruit Bud5p to the axial landmark (Kang et al., 2001; Marston et al., 2001), which, via Rsr1p and Cdc24p as described above, causes accumulation of Cdc42p at the chosen site. Thus, understanding the role of Axl2p in axial budding is essential for understanding how a new growth site is selected and assembled in MATa or cells. Axl2p is a type I membrane glycoprotein with its bulk N-terminal region sticking out into the extracellular space, a single transmembrane domain spanning the plasma membrane, and a C-terminal tail staying in the cytoplasm (Roemer et al., 1996). The extracellular domain of Axl2p consists of four putative cadherin-like motifs (Dickens et al., 2002), whose function remains unknown. The expression of Axl2p peaks in late G1, unlike other components of the axial landmark such as Bud3p and Bud4p, whose expression peaks at S/G2 and M phase, respectively (Cho et al., 1998; Spellman et al., 1998; Lord et al., 2000). In addition, Bud3p and Bud4p localize exclusively to the bud neck in a septin-dependent manner in medium- and large-budded cells (Chant et al., 1995; Sanders and Herskowitz, 1996), whereas Axl2p localizes to the sites of polarized growth early in the cell cycle, which depends on an intact secretory pathway (Powers and Barlowe, 1998, 2002). It is not known whether Axl2p plays an additional role in polarized growth independently of its role in bud-site selection, as suggested by its early cortical localization. Axl2p also localizes to the bud neck some time after bud emergence (Halme et al., 1996; Roemer et al., 1996), but it is not clear whether Bud3p and Bud4p are required for this neck localization (Halme et al., 1996; Roemer et al., 1996). It is also not known how Axl2p interacts with Ax1lp, Bud3p, and Bud4p to form a functional axial landmark.

The septins are a family of GTP-binding proteins that form heteromeric complexes and filaments in vivo and in vitro (Gladfelter et al., 2001b; Longtine and Bi, 2003; Hall and Russell, 2004; Joo et al., 2005; Versele and Thorner, 2005; Kinoshita, 2006). Septins play important roles in cytokinesis, mitosis, protein trafficking, and apoptosis. Overexpression of septins has been associated with variety of different human tumors (Hall and Russell, 2004; Hall et al., 2005). In S. cerevisiae, septins localize to the bud neck throughout the cell cycle, and they function as a scaffold, upon which proteins involved in axial budding such as Bud3p and Bud4p, chitin synthesis, and cytokinesis are anchored (Gladfelter et al., 2001b; Longtine and Bi, 2003; Versele and Thorner, 2005). Septins also function as a diffusion barrier at the bud neck to prevent the free movement of membrane or membrane-associated proteins between cellular compartments (Barral et al., 2000; Takizawa et al., 2000; Dobbelaere and Barral, 2004). Septins undergo dynamic structural changes in the cell cycle (Caviston et al., 2003; Dobbelaere et al., 2003). At the beginning of the cell cycle, septins are dynamic and freely exchangeable with cytosolic septins at the presumptive bud site. After bud emergence, septin collar or hourglass at the bud neck becomes very stable. After the septin collar is split into two distinct rings in anaphase, septins become dynamic again. These organizational changes presumably underlie the involvements of the septins in different cellular processes such as bud morphogenesis and other functions described above.

The initial formation of the septin collar involves at least three steps: the recruitment of the septins to the presumptive bud site, which depends on Cdc42p and, at least in part, its effectors Gic1p and Gic2p (Caviston et al., 2003; Longtine and Bi, 2003; Iwase et al., 2006); the septin-ring assembly, which depends on Cdc42p GTP hydrolysis (Gladfelter et al., 2002; Smith et al., 2002; Caviston et al., 2003) and its effector, the PAK Cla4p (Cvrckova et al., 1995; Weiss et al., 2000; Goehring et al., 2003; Gladfelter et al., 2004; Kadota et al., 2004; Versele and Thorner, 2004); and the septin-ring maturation into a collar or hourglass, which is regulated by a number of proteins, such as Bni5p (Lee et al., 2002), Nap1p (Mortensen et al., 2002; Gladfelter et al., 2004; Iwase and Toh-e, 2004), and the protein kinases Cla4p (Dobbelaere et al., 2003; Versele and Thorner, 2004), Gin4p (Longtine et al., 1998a) and Elm1p (Bouquin et al., 2000). Many more septin regulators are likely to be identified if more sensitive approaches are used. For example, overexpression of the Swe1p kinase, which causes a G2 delay and elongated bud morphology by inhibiting Cdc28p kinase activity, is known to exacerbate a septin defect (Gladfelter et al., 2005). By using this tool, a number of proteins, including Bud3p, Bud4p, and Bud5p, are also found to play a role in septin organization, although the timing and the mechanisms underlying the functions of these proteins in septin organization remain unknown.

Many mutations in the effector-loop region of CDC42 cause septin defects to some extent (Richman et al., 1999; Kozminski et al., 2000; Richman and Johnson, 2000; Gladfelter et al., 2001a). Two such mutations, cdc42^V36T and cdc42^V36A, were found to cause a decrease in the intrinsic GTPase activity of Cdc42p, elongated bud morphology, and septin-organization defects (Gladfelter et al., 2001a, 2002). A similar mutation, cdc42^V36G, was identified from our screen for temperature-sensitive mutations by a polymerase chain reaction (PCR)-based random mutagenesis approach and was found to display a highly penetrant defect in septin organization even at the permissive temperature (Caviston et al., 2003). We used this allele to perform gene dosage-suppressor screens in the hope of identifying genes that would help us elucidate how Cdc42p regulates septin organization. AXL2 and several other functionally relevant genes were identified from such screens. Detailed structure–function analysis indicates that Axl2p is recruited to the bud neck in a Bud3p- and Bud4p-dependent manner via an interaction between the cytoplasmic tail of Axl2p and Bud4p. The role of Axl2p in the sequential assembly of the axial landmark and in the selection of the bud site does not require its expression in late G1. In addition, we found that Axl2p interacts with Cdc42p and other polarity-establishment proteins and that it regulates septin organization early in the cell cycle independently of its role in bud-site selection.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Strains and Genetic Methods
Yeast strains used in this study are listed in Table 1. Standard culture media and genetic techniques were used except where noted (Guthrie and Fink, 1991). Escherichia coli strains DH12S (Invitrogen, Carlsbad, CA) and BL21 (Invitrogen) were used as hosts for plasmid manipulation and recombinant protein expression, respectively. Oligonucleotide primers for PCR were purchased from Integrated DNA Technologies (Coralville, IA).

YEp352-BUD3 was constructed by inserting an 8.2-kb PstI fragment carrying BUD3 from p35-1 (Chant et al., 1995) into PstI-digested YEp352. YEp352-BUD4 was constructed by inserting a 7.3-kb ScaI fragment carrying BUD4 from pJC109 (kindly provided by M. Lord, University of Vermont) into SmaI-digested YEp352. YIp128-CDC3-GFP was constructed by inserting a 5.3-kb SalI–EcoRI fragment carrying CDC3-GFP from YIp211-CDC3-GFP into YIplac128. YIplac128-CDC3-mCherry was constructed by PCR-amplifying a NotI fragment carrying mCherry by using a pair of primers with NotI site and the plasmid pKT355 (supplied by K. Thorn, University of California at San Francisco) as the template, the resulting NotI–mCherry fragment was used to replace the NotI-GFP fragment in YIp128-CDC3-GFP. YIp211-CDC3-GFP, YIp128-CDC3-GFP, and YIplac128-CDC3-mCherry were linearized with BglII for integration at the CDC3 locus in yeast. Plasmids pUG23-AXL2, pUG23-AXL2-N (residues 1–503), pUG23-AXL2-N+TM (residues 1–535), and pUG23-AXL2-C (residues 530–823) were constructed via gap repair of PCR-amplified AXL2 fragments onto EcoRI-digested pUG23 vector in yeast. Yeast chromosomal DNA carrying wild-type AXL2 was used as the template for the PCRs, and the AXL2 fragments were inserted between the PstI and EcoRV sites of the pUG23 vector.

AXL2-C and various regions of AXL2-C in Figure 4B of this study were first amplified by PCR as EcoRI–BamHI fragments, and they were inserted into EcoRI- and BamHI-digested pGBDU-C1 vector. These fragments in pGBDU-C1 were further digested with EcoRI and SalI and ligated into the pUG36 vector to generate GFP fusions to various regions of Axl2p-C. To generate a series of strains each containing a modified allele of AXL2 at its chromosomal locus, several strategies were used. To generate chromosomal axl2^1-823, axl2^1-725, axl2^1-685, axl2^1-646, and axl2^1-627 constructs (1-823 stands for residues 1–823 of Axl2p) in yeast, AXL2-C^530-823, AXL2-C^530-725, AXL2-C^530-685, AXL2-C^530-646, and AXL2-C^530-627 were isolated as EcoRI–SalI fragments from pGBDU-based constructs and ligated into EcoRI- and SalI-digested pRS306-T_CYC1 vector, which was constructed by inserting a 297-base pair EcoRI–KpnI fragment containing the 3'-UTR region of CYC1 gene from the pUG36 vector into the pRS306 vector. The resulting plasmids were linearized with SphI for integration at the AXL2 locus in yeast. After integration, wild-type AXL2 on the chromosome will be replaced by these mutant constructs. To generate chromosomal axl2^{1-544, 726-823}, axl2^{1-544, 686-823}, axl2^{1-544, 628-823}, axl2^{1-544, 641-725}, axl2^{1-544, 641-685}, and axl2^{1-544, 686-725} constructs in yeast, AXL2-C^726-823, AXL2-C^686-823, AXL2-C^628-823, AXL2-C^641-725, AXL2-C^641-685, and AXL2-C^686-725 were isolated as EcoRI–SalI fragments from pGBDU-based constructs and ligated into EcoRI- and SalI-digested pRS306-AXL2-MidT_CYC1 vector, which was constructed by inserting a PCR-amplified BamHI–EcoRI AXL2 fragment encoding residues 253–544 into the pRS306-T_CYC1 vector. The resulting plasmids were linearized with HindIII for integration at the AXL2 locus in yeast. After integration, wild-type AXL2 on the chromosome will be replaced by these mutant constructs. To generate the chromosomal axl2^1-544 construct in yeast, a PCR-amplified SphI–BamHI fragment containing a stop codon and the 3'-untranslated region (UTR) (407 base pairs) of AXL2 right after the SphI site was inserted into SphI- and BamHI-digested pRS426-AXL2 vector to generate pRS426-AXL2^1-544. The purpose of this step is to introduce a stop codon and the 3'-UTR after the internal SphI site of AXL2. The AXL2 fragment encoding residues 239–544 with the stop codon and the 3'-UTR was isolated as a 1.4-kb EcoRI–BamHI fragment from pRS426-AXL2^1-544 and was inserted into the pRS306 vector to generate pRS306-AXL2-MS1, which was then linearized with HpaI to integrate at the AXL2 locus in yeast.

Plasmid pRS426-AXL2 was constructed by inserting a PCR-amplified SacI–BamHI fragment carrying AXL2, including 970 base pairs of upstream and 407 base pairs of downstream sequences, into the SacI- and BamHI-digested pRS426 vector. Plasmid pRS426-AXL2^1-646 was constructed by replacing the SphI–KpnI insert of pRS426-AXL2 with the SphI–KpnI fragment from pRS306-AXL2^530-646. Plasmids pRS426-AXL2^{1-544, 641-725}, pRS426-AXL2^{1-544, 641-685}, and pRS426- AXL2^{1-544, 726-823} were constructed by replacing the HindIII–KpnI insert of pRS426-AXL2 with the HindIII–KpnI fragments from pRS306-AXL2^{253-544, 641-725}, pRS306-AXL2^{253-544, 641-685}, and pRS306-AXL2^{253-544, 726-823}, respectively. Plasmid pRS425-AXL2-3HA was constructed by inserting a 4.9-kb SpeI fragment of AXL2-3HA from pJC246 (Lord et al., 2000) into SpeI-digested pRS425 vector (2µ, LEU2).

The generation of BUD3-C-GFP and BUD3-AXL2-C-GFP fusion constructs involves several steps. First, a PCR-amplified BamHI–EcoRI fragment encoding residues 1477–1636 of Bud3p without the stop codon was inserted into the pRS306-T_CYC1 vector to generate pRS306-BUD3-C-T_CYC1. Second, an EcoRI–SalI fragment carrying AXL2-C, encoding residues 530–823 of Axl2p, from pGBDU-AXL2-C was inserted into pRS306-BUD3-C-T_CYC1 to generate pRS306-BUD3-AXL2-C. Third, a unique MluI site within T_CYC1 of pUG35 and pUG23-AXL2-C was destroyed by MluI digestion followed by filling-in with the Klenow fragment of DNA polymerase I to generate pUG35-MluI and pUG23-AXL2-C-MluI. Fourth, the EcoRI–KpnI fragment of yEGFP3-T_CYC1 from pUG35-MluI was inserted into EcoRI- and KpnI-digested pRS306-BUD3-C-T_CYC1 to generate pRS306-BUD3-C-GFP. Similarly, the SphI–KpnI fragment of AXL2-C-yEGFP3-T_CYC1 from pUG23-AXL2-C-MluI was inserted into SphI- and KpnI-digested pRS306-BUD3-AXL2-C vector to generate pRS306-BUD3-AXL2-C-GFP. Plasmids pRS306-BUD3-C-GFP and pRS306-BUD3-AXL2-C-GFP were linearized with MluI and integrated at the BUD3 locus of LSY134 (a axl2::KanMX) to generate strains JGY1498 and JGY1499, respectively. Similarly, these digested plasmids were integrated at the BUD3 locus of LSY143 (a axl2::KanMX gin4::TRP1) to generate strains JGY1612 and JGY1613, respectively.

To generate chromosomal GIC2-C-GFP and GIC2-AXL2-C-GFP fusion, a 0.93-kb SpeI and EcoRI fragment carrying GIC2-C, which encodes residues 75–383 of Gic2p, was amplified by PCR, and it was ligated into SpeI- and EcoRI-digested pRS306-BUD3-C-GFP and pRS306-BUD3-AXL2-C-GFP vectors to replace BUD3-C with GIC2-C fragment. The resulting plasmids, pRS306-GIC2-C-GFP and pRS306-GIC2-AXL2-C-GFP, were digested with AflII and integrated at the GIC2 locus of LSY134 (a axl2::KanMX) to generate strains JGY1624 and JGY1625, respectively. Similarly, these digested plasmids were integrated at the GIC2 locus of LSY143 (a axl2::KanMX gin4::TRP1) to generate strains JGY1626 and JGY1627, respectively.

Plasmids pEGKT-CDC42, pEGKT-BEM1, pEGKT-CDC24, pEGKT-RHO1, and pEGKT-GAB1 expressed N-terminal glutathione S-transferase (GST) fusion proteins under the control of the GAL1 promoter of pEGKT (2µ, URA3) (Mitchell et al., 1993). To generate pEGKT-CDC42^Q61L, pEGKT-CDC42^T17N, and pEGKT-CDC42^D57Y, a 0.6-kb BamHI–HindIII fragment carrying CDC42^Q61L, CDC42^T17N, and CDC42^D57Y, respectively, from pGEX-KG-CDC42^Q61L, pGEX-KG-CDC42^T17N, and pGEX-KG-CDC42^D57Y (supplied by D. Lew, Duke University Medical Center) (Gladfelter et al., 2002), was inserted into pEGKT vector. To generate pGEX-KG-CDC42^D57Y, a 0.6-kb EcoRI–SacI fragment carrying CDC42^D57Y from pHB2-GST-CDC42^D57Y (Moskow et al., 2000) was used to replace the EcoRI–SacI fragment carrying CDC42 in pGEX-KG-CDC42 (Gladfelter et al., 2002). Plasmid pEGKT-AXL2-C was constructed by inserting the BamHI–SalI fragment carrying AXL2-C from pUG36-AXL2-C into pEGKT vector. To construct the pEGKT306 vector (integrative, URA3), an 2.0-kb NcoI–BamHI fragment containing the GAL1 promoter and GST gene from pEGKT-AXL2-C was ligated into the NcoI and BamHI sites of pRS306-T_CYC1 vector. Various regions of AXL2-C in the EcoRI–SalI fragments from the pGBDU-based constructs were inserted into pEGKT306 to generate pEGKT306-AXL2-C and its derivatives carrying different regions of AXL2-C. The resulting constructs were linearized with NcoI to integrate at the ura3 locus in yeast.

Complete deletion of AXL1 was constructed in diploid strain YEF473 (Bi and Pringle, 1996) by using the PCR-based method with pFA6a-KanMX6 template (Longtine et al., 1998b). The heterozygous diploid strain was sporulated, and tetrads were dissected to generate strain JGY841. To construct strains LSY134 (a axl2::KanMX) and LSY136 ( axl2::KanMX), the axl2:: KanMX locus in strain YEF3490 (Invitrogen) was amplified by PCR and transformed into YEF473. The resulting heterozygote was sporulated, and the tetrads were dissected to generate LSY134 and LSY136. Strains LSY220 and LSY222 were constructed by crosses of LSY134 with KNY311 ( BUD3::GFP-TRP1) and KNY313 ( BUD4::GFP-TRP1) strains (supplied by J. Pringle, Stanford University Medical Center), followed by tetrad dissection. Similarly, JGY1437 and JGY1439 were constructed by crosses of LSY220 and LSY222 with AM101 and KNY118, followed by tetrad dissection. Strains JGY860, JGY1061, JGY1063, and JGY1065 were constructed by crosses of YEF2921 carrying plasmid pRS316-CDC42 with strains M-1077 (Longtine et al., 2000), JGY841, AM101, and KNY118, respectively, and followed by loss of pRS316-CDC42 and tetrad dissection. To generate JGY864, strains M-1075 and LSY136 were crossed, and tetrads of heterozygous diploid were dissected to generate JGY808 ( axl2::KanMX swe1::LEU2), which was then crossed with YEF2921 carrying pRS316-CDC42, and tetrads of heterozygous diploid were dissected after the loss of pRS316-CDC42 to generate JGY864. To construct LSY295 ( axl2::NatMX) for synthetic genetic array (SGA) analysis, the axl2::KanMX locus in strain LSY134 was first switched to axl2::NatMX as described previously (Tong et al., 2001), which was then PCR amplified and transformed into the strain Y5563 to yield LSY295. LSY295 was crossed to ordered arrays of yeast deletion strains, which include deletions of all nonessential genes in S. cerevisiae, to identify genes that display synthetic-lethal or synthetic-sick relationship with AXL2 (Tong et al., 2001).

To construct JGY1382, chromosomal BUD3 in the strain ML550 (a BUD4::13MYC-TRP1) (Lord et al., 2002) was first tagged with 3HA epitope by the PCR-based method with pFA6a-3HA-His3MX template (Longtine et al., 1998b) to generate JGY1332 (a BUD3::3HA-HIS3MX BUD4::13MYC-TRP1), which was then crossed with LSY136, and tetrads of the heterozygous diploid were dissected to generate JGY1382.

Genetic Suppressor Screens
Genetic suppressor screens were carried out in the yeast strain YEF2921 (a cdc42^V36G). Transformants were grown on SC-Ura plates at 35°C. In the first screen, a YEp24 (2µ, URA3)-based genomic DNA library (Carlson and Botstein, 1982) was used. From 6600 transformants screened, 52 suppressors were isolated; 41 of them were determined to carry CDC42 (7x), BEM1 (12x), CLA4 (6x), ELM1 (9x), and AXL2 (7x). The rest suppressor plasmids each carry multiple open reading frames (ORFs) without clearly defined roles in cellular morphogenesis. Among the seven AXL2-containing suppressor plasmids, four carry full-length AXL2, which encodes a protein of 823 amino acids, whereas the other three plasmids carry two different 3'-end–truncated versions of AXL2, encoding residues 1–621 and 1–685 of Axl2p, respectively. The truncations occur at the two Sau3AI restriction sites near the 3' end of the AXL2 ORF. Subcloning analysis confirmed that AXL2 was solely responsible for the suppression. In the second screen, a p366 (CEN, LEU2)-based genomic DNA library was used (Bi and Pringle, 1996). From 6000 transformants, 33 suppressors were isolated, and 20 of them were determined to carry CDC42 (11x), BEM1 (1x), CLA4 (2x), and AXL2 (6x).

Microscopy
A computer-controlled Eclipse 800 microscope (Nikon, Tokyo, Japan) and a high-resolution charge-coupled device camera (model C4742-95; Hamamatsu Photonics, Bridgewater, NJ) were used to visualize cell morphology and GFP-tagged proteins by differential interference contrast (DIC) and fluorescence microscopy, respectively. The images were acquired using Phase 3 Imaging Systems (Glen Mills, PA). To determine the budding pattern of a yeast strain, bud scars were visualized by staining with calcofluor as described previously (Pringle, 1991).

GST Pull-Down Assay
Yeast strains carrying pEGKT vector or pEGKT-AXL2-C were grown at 30°C to A₆₀₀ of 0.40.5 in SC-Ura medium containing 2% raffinose. Galactose was added to a final concentration of 2%, and the cultures were grown for another 4 h to induce the expression of GST fusion proteins. Cells were then harvested and washed with phosphate-buffered saline (PBS) buffer. About 60 A₆₀₀ units of cells were resuspended in 0.5 ml of cell lysis buffer (PBS buffer, 1 mM EDTA, 0.1% NP-40, plus a cocktail of protease inhibitors), and they were disrupted by vortexing in the presence of glass beads at 4°C. Cell debris was removed by centrifugation at 10,000 x g for 15 min. The supernatant was incubated with 50 µl of glutathione-Sepharose 4B beads (GE Healthcare, Little Chalfont, Buckinghamshire, UK) with gentle rocking at 4°C for 16 h. The beads were washed three times with cell lysis buffer, and the bound proteins were eluted by boiling in 100 µl of 2X SDS sample buffer for 5 min. Samples were then analyzed by SDS-polyacrylamide gel electrophoresis (PAGE), followed by standard immunoblotting procedure by using enhanced chemiluminescence reagents. Primary antibodies used were mouse monoclonal antibodies against GST, MYC, hemagglutinin (HA), and GFP (Covance Research Products, Richmond, CA). Secondary antibody was horseradish peroxidase-conjugated rabbit anti-mouse immunoglobulin G (IgG) (Jackson ImmunoResearch Laboratories, West Grove, PA). GST fusion proteins in the pull-down precipitates were analyzed by 12.5% SDS-PAGE and immunoblotted with an anti-GST antibody.

For detecting the interactions of Axl2p with Cdc42p, Bem1p, and Cdc24p, strain JGY1570 carrying pEGKT-based constructs was grown in liquid SC-Leu-Ura media containing 2% raffinose. Cells were harvested after galactose induction for 4–8 h and lysed in 0.5 ml of GPLB buffer (PBS buffer, 5 mM MgCl₂, 1 mM dithiothreitol, plus a cocktail of protease inhibitors). Triton X-100 was added to a final concentration of 0.5% to the cell lysates. Cell lysates were gently rocked at 4°C for 60 min to extract membrane-bound proteins, such as Axl2p and Cdc42p. After centrifugation at 10,000 x g for 15 min, the supernatants were used to perform the pull-down assay.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Identification of AXL2 as a Dosage Suppressor of a cdc42^V36G Mutant
Previously, we have shown that a temperature-sensitive (Ts) cdc42^V36G mutant displayed severe defects in septin organization and cell morphology, including the formation of elongated buds and of cell chains at the permissive temperature of 23°C (Caviston et al., 2003). The mutated residue valine 36 sits in the "effector loop" region of Cdc42p, which is thought to mediate the interactions of Cdc42p with its downstream effectors (Wittinghofer and Nassar, 1996) and its regulators such as GEFs, GAPs, and GDIs (DerMardirossian and Bokoch, 2005). Because cdc42^V36G cells display highly penetrant septin-organization defects (Caviston et al., 2003) (Figure 1B), we thought that by isolating dosage suppressors of cdc42^V36G, we might be able to gain some new insight into how Cdc42p is involved in septin organization. We performed two genetic screens to search for genes that, when overexpressed, could rescue the growth defect of cdc42^V36G mutant at 35°C (see details of the screens in Materials and Methods). BEM1, CLA4, and AXL2 were isolated from both the multicopy-suppressor and the low copy-suppressor screens, whereas ELM1 was isolated only from the multicopy-suppressor screen (Figure 1A; data not shown). Among these suppressors, BEM1 showed the strongest suppression for the growth defect. CLA4 and ELM1 were weaker than AXL2. The morphological defects of the cells at 24°C also seemed to be suppressed to some extent by these multicopy suppressors with CLA4 and BEM1 showing the best suppression (data not shown).

View larger version (67K): [in this window] [in a new window]   Figure 1. Multicopy AXL2 suppresses cdc42V36G in the absence of SWE1. (A) Suppression of cdc42V36G by multicopy BEM1, AXL2, ELM1, and CLA4. Strain YEF2921 (a cdc42V36G) carrying the YEp24 vector, YEp24-CDC42, YEp24-BEM1, YEp24-AXL2, YEp24-ELM1, or YEp24-CLA4 was streaked onto SC-Ura plates and incubated for 3 d at the indicated temperatures. (B) Deletion of SWE1 suppresses the morphological, and, to some extent, the septin-organization defects of cdc42V36G cells. Strains YEF2921 (a cdc42V36G) and JGY860 (a cdc42V36G swe1) carrying YIp211-CDC3-GFP were grown at 24°C. Cell morphology and septin organization, indicated by Cdc3p-GFP, were visualized by DIC and fluorescence microscopy. (C) Differential suppression of cdc42V36G swe1 mutant by multicopy BEM1, AXL2, ELM1, and CLA4. Strain JGY860 (a cdc42V36G swe1) carrying the YEp24 vector, YEp24-CDC42, YEp24-BEM1, YEp24-AXL2, YEp24-ELM1, or YEp24-CLA4 was streaked onto SC-Ura plates and incubated at 24°C or 35°C for 3 d. (D) SWE1 modulates the genetic interaction between CDC42 and AXL2. Strains M-1075 (a swe1), JGY860 (a cdc42V36G swe1), and JGY864 (a cdc42V36G axl2 swe1) were transformed with equal amounts of plasmid pRS316 or pRS316-SWE1-12MYC and grown on SC-Ura plate at 24°C for 3 d.

Unlike BEM1 and CLA4, previous studies had not identified a genetic interaction of CDC42 with either AXL2 or ELM1. Neither AXL2 nor ELM1 was isolated from our previous multicopy-suppressor screens by using the cdc42-201 and cdc42^G60D mutants (Gao et al., 2004). AXL2 encodes a type I transmembrane glycoprotein that participates in axial bud-site selection in haploid MATa or cells (Halme et al., 1996; Roemer et al., 1996). ELM1 encodes a protein kinase that is involved in septin organization and entry into mitosis (Blacketer et al., 1993; Sreenivasan and Kellogg, 1999; Bouquin et al., 2000). Because Elm1p plays an important role in the phosphorylation of Cla4p and of a Nim1-related protein kinase, Gin4p (Sreenivasan and Kellogg, 1999; Bouquin et al., 2000) and because Gin4p is directly involved in the assembly of the septin ring (Longtine et al., 1998a; Mortensen et al., 2002), we examined whether overexpression of GIN4 suppresses the cdc42^V36G mutant. We found that a low-copy plasmid carrying GIN4 indeed suppressed the cdc42^V36G mutant at 35°C (data not shown). The suppression of the cdc42^V36G mutant by ELM1, CLA4, and GIN4, which encode protein kinases that function coordinately to control septin organization (Tjandra et al., 1998; Bouquin et al., 2000), is consistent with the fact that the most prominent defect of the cdc42^V36G mutant is in the assembly of the septin ring (Caviston et al., 2003) (Figure 1B).

AXL2 Suppresses the cdc42^V36G Mutant Independently of SWE1
Defects in bud formation, particularly in septin organization, in a number of mutants, including elm1, cla4, and gin4 mutants, trigger a Swe1p-dependent G2 delay that inhibits the apical-to-isotropic switch of bud growth, leading to an elongated bud morphology (Barral et al., 1999; Bouquin et al., 2000; Longtine et al., 2000; Weiss et al., 2000; Gladfelter et al., 2005). To investigate whether the elongated bud morphology of cdc42^V36G cells depends on Swe1p, we constructed a cdc42^V36G swe1 strain and found that deletion of SWE1 dramatically improved the morphology of cdc42^V36G cells. The elongated bud morphology disappeared, and the septin defects, in particular, the localization of septins to the bud cortex, was greatly reduced (Figure 1B). However, septins at the bud neck were still misorganized. Cells still formed clusters and displayed temperature-sensitive growth at 35°C (Figure 1C). These results indicate that Swe1p plays a role in the exacerbation of the septin defects of the cdc42^V36G cells, supporting a conclusion reached in a recent study (Gladfelter et al., 2005). However, it is noteworthy that the septin defects at the bud neck of the cdc42^V36G cells are largely independent of Swe1p.

The phenotypes of the cdc42^V36G swe1 mutant, including its temperature-sensitive growth and its improved cell morphology and septin organization, offered a potential way to classify our isolated suppressors into distinct functional groups. Indeed, we found that multicopy BEM1 or AXL2 suppressed the cdc42^V36G swe1 mutant at 35°C, whereas multicopy CLA4 or ELM1 failed to do so (Figure 1C). These results suggest that the suppression by CLA4 or ELM1 may be mediated through septin organization and/or Swe1p. In contrast, the suppression by AXL2 and BEM1 may occur via other unknown mechanisms.

To probe the functional interaction between CDC42 and AXL2 further, we found that cdc42^V36G and axl2 were synthetically sick at 24°C. Tetrad analysis showed that the majority of the cdc42^V36G axl2 double mutants failed to form colonies and that the few viable double mutants grew much more slowly than the cdc42^V36G single mutants did at 24°C. In contrast, the cdc42^V36G axl2 swe1 triple mutants from the same tetrad analysis grew very well (Figure 1D). Because of the heterogeneous behavior by the different cdc42^V36G axl2 segregants in their abilities to form colonies, it was difficult to determine the nature of the synthetic-sickness observed between cdc42^V36G and axl2. This problem was circumvented by introducing a centromere-based plasmid carrying SWE1 into the relatively healthy cdc42^V36G axl2 swe1 triple mutant. The SWE1 plasmid inhibited the growth of the triple mutant as well as the cdc42^V36G swe1 double mutant, albeit to a lesser extent (Figure 1D). These results suggest that the synthetic-sickness between cdc42^V36G and axl2 is likely due to an exacerbated septin-organization defect that triggers Swe1p-dependent cell cycle delay (it is nearly impossible to distinguish the septin defects quantitatively in cdc42^V36G cells versus in cdc42^V36G axl2 cells, due to the already severe and diverse septin defects in the single mutant). These results also implicate a role of Axl2p in septin organization, a point that is strongly supported by other genetic studies described below.

Axl2p Interacts with Cdc42p, Bem1p, and Cdc24p In Vivo
How did the multicopy AXL2 suppress the cdc42^V36G mutant? One possibility is that Axl2p directly interacts with Cdc42p and/or its regulators to regulate Cdc42p functions. To examine this possibility, we performed GST pull-down assays. GST-Cdc42p, GST-Bem1p, or GST-Cdc24p, along with controls, were expressed in axl2 cells carrying AXL2-3HA on a high-copy plasmid, and they were pulled down with glutathione-Sepharose beads. The presence of Axl2p-3HA was detected by immunoblotting with an anti-HA antibody. Axl2p-3HA was pulled down with GST-Cdc42p, GST-Bem1p, and also weakly with GST-Cdc24p, but not with the two controls, GST-Rho1p and an endoplasmic reticulum protein GST-Gab1p (Grimme et al., 2004) (Figure 2A), indicating that a pool of Axl2p forms complexes with Cdc42p, Bem1p, and Cdc24p in vivo. Because Cdc42p and Bem1p are known to interact with each other (Gulli et al., 2000; Bose et al., 2001), we examined whether Axl2p could still interact with Cdc42p in the absence of Bem1p. As shown in Figure 2B, GST-Cdc42p still pulled down Axl2p-3HA, suggesting that Axl2p-Cdc42p interaction can occur independently of Bem1p.

View larger version (57K): [in this window] [in a new window]   Figure 2. Genetic and biochemical interactions between Axl2p and Cdc42p. (A) Axl2p forms complex with Cdc42p, Bem1p, and Cdc24p in vivo. Strain JGY1570 (a axl2, pRS425-AXL2-3HA) carrying the pEGKT vector (GST), pEGKT-CDC42 (GST-Cdc42p), pEGKT-BEM1 (GST-Bem1p), pEGKT-CDC24 (GST-Cdc24p), pEGKT-RHO1 (GST-Rho1p), or pEGKT-GAB1 (GST-Gab1p) was grown to the exponential phase in liquid SC-Leu-Ura medium containing 2% raffinose at 30°C. Galactose was added to a final concentration of 2%, and the cultures were grown for another 4 h to induce the expression of GST fusion proteins. GST-tagged proteins were pulled down with glutathione-Sepharose beads from equal amounts of Triton X-100–solubilized cell lysates. The amount of soluble Axl2p-HA in the cell lysates (Input) and in the pull-down precipitates (Bound) was analyzed by SDS-PAGE and immunoblotted with an anti-HA antibody. The amount of GST fusion proteins in the pull-down precipitates (Bound) was immunoblotted with an anti-GST antibody. (B) Axl2p interacts with Cdc42p independently of Bem1p. The GST-pull down experiment was performed as described in A except that the induction time for the expression of GST or GST-Cdc42p by galactose in host strain YEF4889 (bem1) was 8 h instead of 4 h for all other strains, due to the slow growth of the bem1 strain. (C) Axl2p preferentially interacts with GDP-bound form of Cdc42p in vivo. A similar experiment, as described in A, was performed in strain JGY1570 (a axl2, pRS425-AXL2-3HA) carrying pEGKT-CDC42 (WT), pEGKT-CDC42Q61L, pEGKT-CDC42T17N, or pEGKT-CDC42D57Y. (D) The suppression of the cdc42V36G mutant depends on the first one third of Axl2p-C. Strain YEF2921 (a cdc42V36G) carrying the pRS426 vector, pRS426-AXL2 (1-823), pRS426-AXL21-544, pRS426-AXL21-646, pRS426-AXL21-544, 641-725, pRS426-AXL21-544, 641-685, and pRS426-AXL21-544, 726-823 was streaked onto SC-Ura plates and incubated at 24 or 35°C for 3 d.

In addition to the full-length gene, three truncated versions of AXL2 were isolated from the multicopy-suppressor screen, which are predicted to encode two C-terminally truncated proteins Axl2p^1-621 and Axl2p^1-685. Axl2p^1-621 carries only the first one third of its intracellular domain. To characterize further the region of Axl2p-C (the intracellular domain of Axl2p) required for the suppression, a series of constructs carrying different regions of Axl2p-C fused to the N-terminal extracellular domain plus the transmembrane domain of Axl2p were generated and their ability to suppress the cdc42^V36G mutant was examined. Consistently, Axl2p^1-646, a fusion construct carrying the first one third of Axl2p-C (530-646 amino acids), was sufficient to suppress the cdc42^V36G mutant at 35°C (Figure 2D). In contrast, two other fusion constructs carrying either the middle portion (1-544 plus 641-725 amino acids) or the last one third (1-544 plus 726-823 amino acids) of Axl2p-C were unable to suppress. These results, together with our finding that Axl2p^1-646 lacks the Bud4p-interacting region and likely localizes to the cortex of a small bud, but not to the bud neck of middle- and large-budded cells (Figure 4, A and B), indicate that the first one third of Axl2p-C is involved in an interaction with Cdc42p and that this interaction likely occurs early in the cell cycle.

The Suppression of the cdc42^V36G Mutant by AXL2 Is Independent of Its Role in Bud-Site Selection
Before this study, Axl2p is known only for its role in the selection of an axial-budding site, a function that requires cooperation with three other components of the axial landmark, Bud3p, Bud4p, and Axl1p. To determine whether Axl2p shares its function in the suppression of cdc42^V36G mutant with these landmark proteins, we examined the effect of overexpression of Bud3p, Bud4p, and Axl1p on cdc42^V36G mutant and found that multicopy BUD3, BUD4, AXL1, and STE23 (STE23 encodes a functional homologue of Axl1p in the processing of mating pheromone a-factor; Adames et al., 1995) did not suppress the cdc42^V36G mutant at 35°C (Figure 3A). In addition, BUD3, BUD4, and AXL1 were all dispensable for the suppression of cdc42^V36G mutant by AXL2, as multicopy AXL2 suppressed cdc42^V36G bud3, cdc42^V36G bud4 and cdc42^V36G axl1 mutant at 35°C (Figure 3B). These results indicate that AXL2 does not share the function in the suppression of cdc42^V36G mutant with other components of the axial landmark.

View larger version (58K): [in this window] [in a new window]   Figure 3. Axl2p displays separable functions in cdc42V36G suppression and bud-site selection. (A) Multicopy BUD3, BUD4, AXL1, and STE23 does not suppress cdc42V36G mutant. Strain YEF2921 (a cdc42V36G) carrying the YEp24 vector, YEp24-AXL1, YEp24-STE23, YEp24-AXL2, YEp24-BUD3, or YEp24-BUD4 was streaked onto SC-Ura plates and incubated at 24 or 35°C for 3 d. (B) BUD3, BUD4, and AXL1 are dispensable for the suppression of cdc42V36G mutant by multicopy AXL2. Strains JGY1061 (a cdc42V36G axl1), JGY1063 (a cdc42V36G bud3), and JGY1065 (a cdc42V36G bud4) were transformed with the YEp24 vector or YEp24-AXL2 and the transformants were streaked onto SC-Ura plates and incubated at 24 or 35°C for 3 d. (C) The middle portion of Axl2p-C is required for the role of Axl2p in bud-site selection. A series of AXL2 constructs carrying different regions of its tail were transformed into strain LSY134 (a axl2). All these constructs contain sequences encoding the N-terminal portion and the transmembrane domain of Axl2p. The budding pattern of each strain was determined by staining of bud scars with Calcofluor. One hundred cells with four or more bud scars were counted for each strain.

Roles of the Middle Portion of Axl2p-C in Axial Bud-Site Selection: Interaction With Other Components of the Axial Landmark
Among the components of the axial landmark, Axl2p displays a unique localization pattern in the cell cycle, an early bud-cortex localization similar to that of some polarity-establishment proteins, and a late bud-neck localization similar to that of other axial-landmark proteins. To determine which region of Axl2p is required for its targeting to the bud neck, the location where other axial-landmark proteins reside, a series of GFP fusion constructs containing different AXL2 fragments under the control of a methionine-regulated promoter were made and introduced into axl2 cells for localization studies. We observed that the N-terminal region of Axl2p localized primarily in the vacuole and occasionally was also observed faintly decorating the plasma membrane (Figure 4A). In contrast, the N-terminal region plus the transmembrane domain localized to the bud cortex in small- to medium-budded cells, and to the bud neck region in large-budded cells. However, the latter localization was often seen more spread out at the bud-neck region. Interestingly, we found that the cytoplasmic tail of Axl2p localized to the bud neck as a tight double ring in medium- to large-budded cells, indicating that Axl2p-C plays a role in anchoring Axl2p to the bud neck. In addition, a fraction of Axl2p-C also localized to the nucleus. Further analysis with a constellation of methionine promoter-controlled GFP-AXL2 C-terminal fragments, which expressed the fusion proteins at comparable levels (Figure 4B), indicates that the minimal bud neck-targeting region consists of amino acids 641-685, which is located in the middle portion of Axl2p-C (Figure 4, A and B, Axl2p-C9).

View larger version (46K): [in this window] [in a new window]   Figure 4. The targeting of Axl2p to the bud neck is mediated by an interaction of Axl2p-C with Bud4p. (A) The pUG23 vector carrying full-length Axl2p (residues 1–823), N-terminal portion (Axl2p-N, residues 1–503), N-terminal portion plus transmembrane domain (Axl2p-N+TM, residues 1–535), the C-terminal tail (Axl2p-C, residues 530–823), and the pUG36 vector carrying the C9 fragment (Axl2p-C9, residues 641–685), the minimal bud neck-targeting region of Axl2p, was transformed into LSY134 (a axl2) strain. The localization of GFP fusion proteins was visualized. (B) The expression levels and the targeting capacities of different Axl2p fragments. The pUG36 vector carrying different regions of Axl2p-C was transformed into LSY134 (a axl2) strain. The resulting strains were grown to exponential phase in SC-Ura media at 25°C, and the localizations of different GFP fusion proteins were determined. The same strains were grown in SC-Ura-Met media at 30°C for 17 h to induce the expression of different GFP–Axl2p fusions. The levels of these fusion proteins were determined by Western blot with an antibody against GFP. (C) The plasmid pUG23-AXL2-C was transformed into BY4742 (WT), YEF3491 ( axl1), YEF3492 ( bud3), and YEF3493 ( bud4) strains. The localization of Axl2p-C-GFP was visualized and quantified. For each strain, 50 cells with nuclei positioning at the extreme ends of the mother-daughter axis were counted. One hundred percent of such cells for the wild-type and axl1 strains displayed bud-neck localization, in comparison with 0 and 2% for the bud3 and bud4 strains, respectively. (D) Axl2p interacts with Bud3p and Bud4p. Strain JGY1382 (a axl2 BUD3-3HA BUD4-13MYC) was transformed with pEGKT or pEGKT-AXL2-C. GST and GST-Axl2p-C were pulled down from cell lysates by glutathione-Sepharose beads. Cell lysates (Input) and bound fractions were resolved by 6% SDS-PAGE (for Bud3p-HA and Bud4p-MYC) or 12.5% SDS-PAGE (for GST-fusions) and immunoblotted with antibodies against HA, MYC, and GST. (E) Axl2p-Bud3p interaction is mediated mainly via the middle one third of Axl2p-C. Strain JGY1464 ( axl2 BUD3-3HA KCC4-13MYC) was transformed with pEGKT306 (GST only), or pEGKT306 carrying various fragments of Axl2p-C. GST and GST fusion proteins were pulled down from cell lysates by glutathione-Sepharose beads. Cell lysates (Input) and bound fractions were resolved by 6% SDS-PAGE (for Bud3p-HA) or 12.5% SDS-PAGE (for GST fusions) and immunoblotted with antibodies against HA and GST. Note that GST-Axl2p-C (641-685) fusion failed to pull down Bud3p-HA, even though this fragment of Axl2p was able to localize to the bud neck (see Axl2p-C9 in A). This presumably reflects differences of various Axl2p fragments in neck-targeting efficiency and/or their differential abilities to form stable Axl2p-Bud3p complexes. (F) Axl2p-Bud3p interaction is mediated by Bud4p. Strains LSY220 (a axl2 BUD3-GFP), JGY1437 (a axl2 BUD3-GFP bud4), LSY222 (a axl2 BUD4-GFP), and JGY1439 (a axl2 bud3 BUD4-GFP) were transformed with pEGKT-AXL2-C. Glutathione-Sepharose beads were used to pull down GST-Axl2p-C from cell lysates. GST-Axl2p-C, Bud3p-GFP, and Bud4p-GFP in the bound fraction (Bound) as well as Bud3p-GFP and Bud4p-GFP in the cell lysates (Input) were resolved by SDS-PAGE and detected by immunoblotting with anti-GST and anti-GFP antibodies.

We also found that Axl2p-C9, the minimal region for bud-neck targeting in the Axl2p cytoplasmic tail, was unable to direct axial budding (Figure 3C). However, this region plus additional C-terminal sequence (residues 641–725), representing the middle portion of Axl2p-C, was able to confer axial budding to axl2 cells (Figure 3C). Thus, the middle portion of Axl2p-C may play an additional role in axial budding, in addition to the recruitment of Axl2p to the bud neck. For example, the cytoplasmic tail of Axl2p has been suggested to recruit and/or activate Bud5p, the GEF for Rsr1p (Kang et al., 2001). It is noteworthy that even though the minimal targeting region, Axl2p-C9, localized to the bud neck, the signal at the neck was relatively weaker than the full-length Axl2p-C (Figure 4A). GST-Axl2p-C9 did not pull down detectable amount of Bud3p (Figure 4E). Even a larger fragment of Axl2p-C (residues 641–725), which is sufficient for axial budding, pulled down less Bud3p than the full-length Axl2p-C did. These results suggest that the minimal targeting domain alone cannot target to the bud neck as efficiently as the full-length tail and/or cannot maintain a stable association with Bud3p, presumably via Bud4p.

Axl2p Plays a Role in Septin Organization
The suppression of cdc42^V36G cells, which display pronounced defects in septin organization, by multicopy AXL2, ELM1, and CLA4 suggests that AXL2 may play a role in septin organization. This conclusion is further supported by the fact that a SGA analysis (Tong et al., 2001) with an axl2 mutant against ordered arrays of >4000 yeast deletion strains identified 38 potential synthetic-sick interactions, and further tetrad analysis on these mutants confirmed a synthetic-sick interaction of axl2 with elm1, but not with the rest of these candidates. To explore the potential role of Axl2p in septin organization further, we examined the synthetic effect of axl2 with gin4 and cla4 mutants, which are, primarily if not exclusively, defective in the assembly of the septin ring (Cvrckova et al., 1995; Longtine et al., 1998a; Weiss et al., 2000; Versele and Thorner, 2004). We found that axl2 gin4 and axl2 cla4 mutants displayed synthetic morphological defects, including elongated bud morphology and cell chains, as well as septin-organization defects (Figure 5A; data not shown). The septins, indicated by Cdc3p-GFP, were often mislocalized to the bud cortex and misorganized at the bud neck (Figure 5A), indicating that Axl2p indeed plays a role in septin organization. The phenotype of axl2 gin4 mutant seemed more penetrant than axl2 cla4 mutant (data not shown), but less so than axl2 elm1 mutant (data not shown). axl2 gin4 cells also grew slightly slower than gin4 cells at 30°C.

View larger version (59K): [in this window] [in a new window]   Figure 5. Axl2p is involved in septin organization. (A) Synthetic defects in cell morphology and septin organization between axl2 and gin4. Strains YEF473A (a WT), LSY136 ( axl2), M-267 (a gin4), and LSY143 (a axl2 gin4) carrying YIp128-CDC3-GFP were grown on SC-Leu plates at 24°C for overnight. Cell morphology and Cdc3p-GFP localization were visualized. (B) Overexpression of the cytoplasmic tail of Axl2p causes septin defects. Strain JGY969 (a CDC3-GFP) carrying pEGKT or pEGKT-AXL2-C was streaked onto SC-Ura plate containing 2% galactose and 1% raffinose and grown at 30°C for 16 h. Then, they were visualized for cell morphology and Cdc3p-GFP. (C) A set of AXL2 constructs as indicated was transformed into LSY143 (a axl2 gin4) strain, and the transformants were grown on SC-Trp plates at 30°C overnight and then visualized for cell morphology. Top, rescue of morphological defect. Bottom, representative cell morphology. The strength of rescue was scored from strong (++++) to no rescue (–).

The Role of Axl2p in Septin Organization Is Mediated Largely by the Middle Portion of Axl2p-C
Because the axl2 gin4 cells display severe defects in cell morphology and septin organization, this provides an excellent opportunity to address the requirement of Axl2p-C in septin organization. A series of constructs with different regions of Axl2p-C fused to the N-terminal region plus the transmembrane domain of Axl2p were introduced into axl2 gin4 cells. The construct carrying the middle portion (residues 641–725) of Axl2p-C rescued the morphological defect of axl2 gin4 cells, whereas the construct carrying the last one-third (residues 726–823) of Axl2p-C did not rescue the morphology at all (Figure 5C). Interestingly, the construct carrying the first one third (residues 530–627) of Axl2p-C also seemed to rescue the morphological defect to some extent (Figure 5C). The morphological rescue by Axl2p-C constructs was correlated with improved septin organization (data not shown). Thus, this study identified a major role for the middle portion of Axl2p-C and a minor role for the first one third of Axl2p-C in septin organization.

Late G1 Expression of Axl2p Is Not Required for Bud-Site Selection, but It Is Required for Septin Organization
The expression of AXL2, BUD3, and BUD4 is cell cycle regulated with their expression peaked in late G1, S/G2, and M phase, respectively (Cho et al., 1998; Spellman et al., 1998; Lord et al., 2000). As a transmembrane protein, Axl2p is thought to rely on the highly polarized secretory machinery in late G1 to be delivered to the correct location (Powers and Barlowe, 1998), and the early expression of AXL2 is thought to play a key role in bud-site selection, as constitutive or delayed expression affects the localization and function of Axl2p (Lord et al., 2000). We reasoned that the observed loss of function could result from the mislocalization of Axl2p alone, and thus, the requirement of the early expression of Axl2p for its role in bud-site selection is still an open question. To examine this possibility, we generated a Bud3p-Axl2p-C chimera protein by fusing AXL2-C to the 3' end of the BUD3 ORF and integrated it into the chromosomal BUD3 locus in axl2 cells. Bud3p-Axl2p-C was expressed under the control of the endogenous BUD3 promoter and served as the sole source of Bud3p and Axl2p in the cell. By this approach, the expression of Axl2p-C was delayed to S/G2 phase without a